U.S. patent application number 12/691509 was filed with the patent office on 2010-08-26 for ceiling microphone assembly.
This patent application is currently assigned to TANDBERG TELECOM AS. Invention is credited to Trygve Frederik MARTON.
Application Number | 20100215189 12/691509 |
Document ID | / |
Family ID | 40627517 |
Filed Date | 2010-08-26 |
United States Patent
Application |
20100215189 |
Kind Code |
A1 |
MARTON; Trygve Frederik |
August 26, 2010 |
CEILING MICROPHONE ASSEMBLY
Abstract
A video teleconferencing directional microphone has two surfaces
joined with an angle of 90.degree. relative to each other, a first
omni directional microphone element arranged adjacent to the
intersection between the two surfaces. The ceiling microphone
assembly also includes a second omni directional microphone element
arranged at a predetermined distance (d) from both surfaces. A
subtractor subtracts the output of the first microphone element
from the output of the second microphone element, and the output of
the subtractor is equalized by an equalizer (H.sub.eq) to generate
an equalized output. The surfaces and subtractor generates a
quarter toroid directivity pattern for the ceiling microphone
assembly. The quarter toroid sensitivity pattern increases
sensitivity in the direction of a sound source of interest, but
reduces sensitivity to any sound waves generated by noise sources
at other locations or reverberations.
Inventors: |
MARTON; Trygve Frederik;
(Oslo, NO) |
Correspondence
Address: |
OBLON, SPIVAK, MCCLELLAND MAIER & NEUSTADT, L.L.P.
1940 DUKE STREET
ALEXANDRIA
VA
22314
US
|
Assignee: |
TANDBERG TELECOM AS
Lysaker
NO
|
Family ID: |
40627517 |
Appl. No.: |
12/691509 |
Filed: |
January 21, 2010 |
Current U.S.
Class: |
381/92 ;
381/122 |
Current CPC
Class: |
H04R 1/406 20130101;
H04R 2430/21 20130101 |
Class at
Publication: |
381/92 ;
381/122 |
International
Class: |
H04R 3/00 20060101
H04R003/00; H04R 1/40 20060101 H04R001/40 |
Foreign Application Data
Date |
Code |
Application Number |
Jan 21, 2009 |
NO |
20090325 |
Claims
1. A microphone assembly, comprising: an L-shaped structure formed
by a first planar surface perpendicularly attached to a second
planar surface; a first microphone element disposed at the
intersection of the first and second planar surfaces; a second
microphone element disposed along a line bisecting an angle formed
by the first and second planar surfaces, the second microphone
element being a predetermined distance from both the first and
second planar surfaces; a first subtractor configured to subtract
an output of the first microphone element from an output of the
second microphone element; and a first equalizer configured to
equalize the output of the first subtractor, the first equalizer
having a frequency response of H.sub.eq, wherein a quarter toroid
sensitivity pattern for the microphone assembly is generated by
acoustical interaction of the two planar surfaces with the two
microphone elements and subtraction of the microphone element
outputs.
2. The microphone assembly according to claim 1, wherein the first
and second microphone elements are omni directional
microphones.
3. The microphone assembly according to claim 1, wherein the
microphone assembly is a ceiling microphone.
4. The microphone assembly according to claim 1, wherein the
subtractor includes a signal inverter to invert the output of the
first microphone element and an adder to combine an output the
inverter with the output of the second microphone element to
generate the output of the subtractor.
5. The microphone assembly according to claim 1, wherein the
quarter toroid sensitivity pattern has a minimum sensitivity at
azimuth angles corresponding to 0 degrees and 180 degrees with
respect to the first surface and the second surface.
6. The microphone assembly according to claim 1, wherein a product
of the predetermined distance and the square-root of two is a
maximum of half of a wavelength corresponding to a highest
frequency to be captured by the microphone assembly.
7. The microphone assembly according to claim 1, wherein the
equalizer frequency response H.sub.eq is proportional to 1 .omega.
2 , ##EQU00002## and .omega. is angular frequency.
8. The microphone assembly according to claim 1, wherein the
equalizer frequency response H.sub.eq includes a low-frequency
roll-off at 80 Hz.
9. The microphone assembly according to claim 2, further
comprising: a third microphone element arranged at twice the
predetermined distance with respect to both planar surfaces; a
second subtractor to subtract the output of the first microphone
element from an output of the third microphone element; a second
equalizer filter to equalize an output of the second subtractor,
the second equalizer having a second frequency response HL.sub.eq;
a high pass filter to filter an output of the first equalizer
filter (H.sub.eq); a low pass filter to filter an output of the
second equalizer filter (HL.sub.eq); and an adder to combine an
output of the high pass filter and an output of the low pass
filter.
10. The microphone assembly according to claim 3, wherein
HL.sub.eq=H.sub.eq.
11. A microphone assembly, comprising: an L-shaped structure formed
by a first planar surface perpendicularly attached to a second
planar surface; a bidirectional microphone element having a front
acoustic input port and a rear acoustic input port; a first
waveguide having an output port associated with the rear acoustic
input port of the bidirectional microphone, an input port of the
first waveguide being disposed adjacent to an intersection of the
first and second planar surfaces; a second waveguide having an
output port associated with the front acoustic input port of the
bidirectional microphone element, an input port of the second
waveguide being arranged at a predetermined distance (d) from the
first and second planar surfaces; and a equalizer filter configured
to equalize an output of the bidirectional microphone element, the
equalizer having a frequency response (H.sub.eq), wherein a quarter
toroid sensitivity pattern for the microphone assembly is generated
by acoustical interaction of the two planar surfaces, the two
waveguides and the bidirectional microphone element.
12. The microphone assembly according to claim 11, wherein the
first and second waveguide are of equal dimensions.
13. The microphone assembly according to claim 11, wherein the
microphone assembly sensitivity pattern has a minimum sensitivity
at azimuth angles (.beta.) corresponding to 0 degrees and 180
degrees with respect to a line coinciding with the intersection
between the first and second planar surfaces.
14. The microphone assembly according to claim 11, wherein a
product of the predetermined distance and the square-root of two is
a maximum of half of a wavelength corresponding to a highest
frequency to be captured by the microphone assembly.
15. The microphone assembly according to claim 11, wherein the
equalizer frequency response (H.sub.eq) is proportional to 1
.omega. 2 , ##EQU00003## and .omega. is angular frequency.
16. The microphone assembly according to claim 11, wherein the
equalizer frequency response (H.sub.eq) includes a low-frequency
roll-off of 80 Hz.
17. A method for creating an quarter toroid directivity pattern in
a microphone assembly, comprising: joining two planar surfaces to
form an L-shaped structure; converting audio waves received at a
first microphone into first audio data, the first microphone being
arranged adjacent to an intersection of the two planar surfaces;
converting audio waves received at a second microphone into second
audio data, the second microphone being arranged at a predetermined
distance (d) from each of the two planar surfaces; subtracting, in
a subtractor, the first audio data from the second audio data; and
equalizing, in an equalizer, an output of the subtractor.
18. The method according to claim 17, wherein the quarter toroid
directivity pattern has a minimum sensitivity at azimuth angles
(.beta.) corresponding to 0 degrees and 180 degrees with respect to
a line coinciding with the intersection between the two planar
surfaces.
19. The method according to claim 17, wherein a produce of the
predetermined distance (d) and the square-root of two is a maximum
of half of a wavelength corresponding to a highest frequency to be
captured by the microphone assembly.
20. The method according to claim 17, wherein the equalizer
frequency response (H.sub.eq) is H eq = 1 .omega. 2 , ##EQU00004##
and .omega. is angular frequency.
21. The method according to claim 17, wherein the microphone
assembly is a ceiling microphone.
Description
CROSS-REFERENCE TO RELATED APPLICATIONS
[0001] This application is based on and claims priority to
Norwegian Application No. NO20090325, filed Jan. 21, 2009, the
entire contents of which are incorporated herein by reference. This
application is also related to copending U.S. application Ser. No.
12/637,444, entitled "Toroid Microphone Apparatus" and filed Dec.
14, 2009, and U.S. application Ser. No. 12/645,701, entitled
"Elevated Toroid Microphone Apparatus" and filed on Dec. 23, 2009,
the entire contents of both U.S. applications being incorporated
herein by reference.
BACKGROUND
[0002] A microphone assembly is provided. More specifically, a
ceiling mounted microphone assembly having a sensitivity pattern
that is independent of the microphone's elevation angle. The
microphone maximizes sensitivity in the direction of a sound source
of interest, but minimizes sensitivity to sound from other
directions.
[0003] Teleconferencing systems, which can implement audio-only
teleconferences or video and audio teleconferences, create meetings
between two or more parties that are separately located, such as in
separate rooms. The rooms may be within a same building or in
different buildings, and the difference building can be located in
different cities, countries, continents, etc. Thus,
teleconferencing systems create meetings that would otherwise
require travel of potentially large distances.
[0004] Video teleconferencing systems create virtual meetings by
transmitting both video and audio data, and thus include one or
more microphones to capture sound waves. The microphones convert
sound waves generated in one video teleconferencing room into
electrical impulses for transmission to another video
teleconferencing room. Audio quality is therefore directly
dependent on the positioning of the microphone within the room, the
acoustics of the room, and to the characteristics of the microphone
itself.
[0005] For example, a conventional microphone used to capture sound
from a sound source of interest, such as a person speaking, will
receive direct sound waves, reflected sound waves and reverberant
sound waves from the source. Direct sound waves travel directly to
the microphone without reflection, and are the sound waves intended
to be captured by microphones. The level of direct sound waves is
inversely proportional to the distance between the sound source of
interest and the microphone receiving the sound.
[0006] Reflected sound waves do not travel directly to the
microphone. Instead, they are reflected multiple times by objects
in the room, or the room itself, before reaching the microphone.
For example, sound waves from a sound source of interest may be
reflected by walls, floors, ceilings, chairs, etc. Reflected sounds
waves that propagate less than 50-80 ms (corresponding to a
propagation distance of 17 to 27 meters) before reaching the
microphone are known as "early reflections." Early reflections have
pressure levels approximately equal to those of direct sound waves,
but are delayed in time.
[0007] Early reflections from the sound source of interest may
positively contribute to the audio received by the microphone.
However, they may also distort the audio because the time delay
causes a phase difference between the early reflections and the
direct sound wave that results in cancellation of part of the
frequency components of the direct sound waves. This phenomenon is
known as "comb filtering", and negatively impacts quality.
[0008] Reflections that propagate for more than 50 to 80 ms (17 to
27 meters) are known as "reverberant sound". Reverberant sound
arrives at the microphone from nearly every direction because these
sound waves have reflected many times within the room. Also, their
pressure level is largely independent of microphone-sound-source
distance. Unlike early reflections, reverberant sound always
contributes negatively to audio quality by creating a "distant",
"hollow", and/or "muffled" characteristic.
[0009] The level of distortion cause by reverberant sound is
determined by a ratio of a level of direct sound to a level of
reverberant sound. For example, if the sound source of interest is
very close to the microphone the ratio of direct sound to
reverberant sound is large, and distortion is small. As the sound
source of interest moves away from the microphone the ratio of
direct sound to reverberant sound will decrease, increasing
distortion.
[0010] A distance at which the level of the direct sound equals the
level of the reverberant sound is known as the "room radius", which
can be determined for every room. As a sound source of interest
moves outside of the room radius, reverberant sound dominates and
distortion increases. Conversely, as the sound source moves within
the room radius the direct sound dominates, and distortion
decreases. Therefore, in conventional microphone systems, the sound
source of interest should remain within the room radius to avoid
significant audio distortion.
[0011] Direct sound, reflected sound, and reverberant sound are not
limited to the sound source of interest, and can also be present
for noise sources in a video teleconferencing room. Noise sources
include, for example, fan noise from ventilation systems, cooling
fan noise from electronic equipment, noises from outside of the
video teleconferencing room, and noises made directly on the table
by people writing with pens, setting down cups, table-top computer
keyboard typing, moving chairs, etc. Conventional teleconferencing
system microphones receive direct, reflected and reverberant sound
waves from these noise sources as well, deteriorating audio
quality.
[0012] Further, each noise source has a different dominant
component. For example, cooling fans installed on electrical
equipment and noise originating from outside the video
teleconferencing room primarily contribute noise in the form of
reverberant sound waves. Noise generated directly on the table-top
surface on which the microphone is placed contributes direct sound
waves that travel parallel to the surface of the table. Some noise
sources, such as ventilation systems, can also contribute multiple
noise components, such as direct and reverberant sound waves.
[0013] Conventional microphones may also contribute noise in the
form of an echo. An echo occurs when sound from a loudspeaker used
to reproduce remote party audio is captured by the microphone and
retransmitted to the remote party. Echoes also have direct,
reflected and reverberant sound components, but dominance of one
component over the others is determined by a
loudspeaker-to-microphone distance, which is not always
constant.
[0014] Echoes are conventionally attenuated with echo cancellers,
which are adaptive filters that train to a loudspeaker-microphone
channel response. However, echo cancellers cannot prevent a
microphone from receiving an echo. Instead, echo cancellers merely
attenuate echoes already present in an audio signal. Because of
their adaptive nature, echo cancellers require time to adapt to a
given response, making time-invariant loudspeaker-microphone
channel responses desirable. In practice, however, microphones are
often repositioned during a video teleconference in order to
capture audio from several different sound sources, and
time-invariant loudspeaker-to-microphone channels are difficult to
achieve. Thus, a conventional video teleconferencing system's echo
cancellers are typically required to adapt multiple times.
Moreover, echo cancellers have difficulty attenuating reverberant
sound components, resulting in increased computational complexity
as the level of reverberant echoes increase.
[0015] The above problems are exacerbated when omni directional
microphones are used in video teleconferencing systems. An omni
directional microphone receives audio from all directions with
equal sensitivity, and thus receives direct, reflected and
reverberant sounds from every sound source within the room,
including noise sources. In fact, only noise sources below the
conference table is attenuated because the table functions as a
barrier to sound pressure waves. Though omni directional
microphones are capable of capturing audio from all sound sources
of interest without being repositioned, the resulting audio quality
is poor because of captured noise.
[0016] One way to improve the quality of audio transmitted by a
video teleconferencing system is to use directional microphones.
Unlike omni directional microphones, a directional microphone has
higher sensitivity with respect to certain directions over others,
and inherently filters sound from at least some noise sources. This
improves audio quality relative to an omni directional microphone,
but also requires that a directional microphone be oriented to
align its direction of highest sensitivity ("main axis") toward the
sound source of interest. Therefore, the directional microphone
requires repositioning every time the sound source of interest
changes position.
[0017] Directional microphones having a cardioid sensitivity
pattern or a bidirectional sensitivity pattern are typically used
in video teleconferencing. A microphone having a cardioid
sensitivity has a directivity function given by:
g(.alpha.)=1/2+1/2.times.cos(.alpha.), where .alpha. is the azimuth
angle of a main axis with respect to horizontal. A typical cardioid
microphone has a maximum sensitivity at .alpha.=0.degree. and a
minimum sensitivity at .alpha.=180.degree..
[0018] A bidirectional microphone has a directivity function given
by: g(.alpha.)=cos(.alpha.), where .alpha. is also the azimuth
angle of a main axis with respect to horizontal. This microphone
has a maximum sensitivity for .alpha.=0.degree. and
.alpha.=180.degree., and a minimum sensitivity when
.alpha.=90.degree. and .alpha.=270.degree.. Because both the
cardioid and bidirectional sensitivity patterns depend on the
azimuth angle of the microphone, sensitivity for these microphones
varies horizontally and vertically.
[0019] As discussed above, either a cardioid microphone or a
bidirectional microphone may be used in a video teleconferencing
system to improve audio quality. Placing the cardioid or
bidirectional microphone on a table also improves audio quality
because the table acts as a sound barrier to sound origination
below the table surface, improving the direct to reverberant audio
ratio.
[0020] Microphone sensitivity may also be improved by placing the
microphone directly on the table-top surface because at this level
the microphone receives direct sound waves and sound waves
reflected by the table (i.e. early reflections). The direct sound
waves and reflected sound waves reflected by the table, however,
remain in phase and combine to form a pressure wave that is double
that of the direct sound wave. This effectively increases the
microphone sensitivity is by six decibels, and is commonly referred
to as the "boundary principle."
[0021] Directional microphones still have the drawback of requiring
the sound source of interest to remain located near the main
sensitivity direction of the microphone. When several people take
part in the meeting, the microphone must be continually readjusted
to avoid diminished audio quality. Thus, parties to the video
teleconference must be aware of the sensitivity patterns of the
microphone and adjust the position of the microphone accordingly.
This makes directional microphones difficult to use.
[0022] One conventional method of reducing sensitivity to noise
from the table and to orient all sound sources of interest to the
"line of sight" (i.e. area of heightened sensitivity) of the
microphone is to hang the microphone from the ceiling. Directional
microphones, such as cardioid microphones, are often used in
hanging microphone applications. However, the sensitivity pattern
of hanging directional microphones is less focused than that of
tabletop microphones because hanging microphones lack the shielding
provided by a table surface. The missing table surface prevents
hanging directional microphones from exploiting the boundary
principle described above, and hanging directional microphones have
relatively higher levels of self-noise compared to their tabletop
counterparts. Conventional hanging directional microphones are also
more susceptible to reverberant sound. Hence, conventional hanging
directional microphones are only suitable for short-range use.
SUMMARY
[0023] A microphone assembly according to the present disclosure
includes an L-shaped structure formed by a first planar surface
perpendicularly attached to a second planar surface. A first
microphone element is disposed at the intersection of the first and
second planar surfaces, and a second microphone element disposed
along a line bisecting an angle formed by the first and second
planar surfaces. The second microphone element is a predetermined
distance from both the first and second planar surfaces. A first
subtractor is provided to subtract an output of the first
microphone element from an output of the second microphone element,
and a first equalizer to equalize the output of the first
subtractor is also provided. The first equalizer has a frequency
response of Heq. A quarter toroid sensitivity pattern for the
microphone assembly is generated by acoustical interaction of the
two planar surfaces with the two microphone elements and
subtraction of the microphone element outputs.
BRIEF DESCRIPTION OF THE DRAWINGS
[0024] A more complete appreciation of the invention and many of
the attendant advantages thereof will be readily obtained as the
same becomes better understood by reference to the following
detailed description when considered in connection with the
accompanying drawings. However, the accompanying drawings and their
exemplary depictions do not in any way limit the scope of the
inventions embraced by this specification. The scope of the
inventions embraced by the specification and drawings are defined
by the words of the accompanying claims.
[0025] FIG. 1 is a schematic drawing of a video teleconferencing
system's audio distribution section that includes microphones
according to an exemplary embodiment of the present disclosure;
[0026] FIG. 2a is a schematic drawing of the sensitivity patterns
of a ceiling microphone assembly arranged overhead according to an
exemplary embodiment of the present disclosure;
[0027] FIG. 2b is another schematic drawing of the sensitivity
patterns of a ceiling microphone assembly arranged overhead
according to an exemplary embodiment of the present disclosure;
[0028] FIG. 3 is a schematic drawing of a microphone assembly
according to an exemplary embodiment of the present disclosure;
[0029] FIG. 4 is an equivalent diagram corresponding to the
microphone assembly of FIG. 3;
[0030] FIG. 5 is another equivalent diagram corresponding to the
microphone assembly of FIG. 3;
[0031] FIG. 6 is a sensitivity patterns of a ceiling microphone
assembly according to an exemplary embodiment of the present
disclosure;
[0032] FIG. 7 is a schematic diagram of a ceiling microphone
assembly according to another exemplary embodiment of the present
disclosure;
[0033] FIG. 8 is a schematic diagram of a ceiling microphone
assembly according to a further exemplary embodiment of the present
disclosure; and
[0034] FIG. 9 is a is a schematic drawing of a processor according
to aspects of the exemplary embodiments of the present
disclosure.
DETAILED DESCRIPTION
[0035] In the following, the present advancement is discussed by
describing preferred embodiments with reference to the accompanying
drawings. However, those skilled in the art will recognize other
applications and modifications within the scope of the disclosure
as defined in the enclosed claims.
[0036] FIG. 1 is a schematic representation of an audio portion of
a video teleconferencing system. In FIG. 1, speaker 10a, in room
110a, and speaker 10b, in room 110b, are engaged in a video
teleconference. Rooms 110a and 110b may be physically adjacent to
each other in the same building, or separated by many hundreds or
thousands of miles, and communication link 140 is used to transfer
video and audio data between rooms 110a and 110b.
[0037] The exemplary communication link 140 may be wired, such as a
PSTN telephone system, Wide Area Network (WAN), Local Area Network
(LAN), or Ad-hoc. The exemplary communication link 140 may also be
a wireless, such as a cellular network, WiMax, Wifi, or via
satellite link. Further, the communication link 140 may also be a
combination of the wired and wireless networks.
[0038] Rooms 110a and 110b of FIG. 1 are mirror images of each
other, and contain the same equipment. Of course, those skilled in
the art will recognize that alternative configurations are embodied
by the advancements described herein. Each room 110a and 110b
includes a ceiling microphone assembly 20a or 20b, a microphone
amplifier 30a or 30b, an A/D converter 40a or 40b, an echo
canceller 50a or 50b, an encoder 60a or 60b, a decoder 70a or 70b,
a D/A converter 80a or 80b, a power amplifier 90a or 90b, and a
loudspeaker 100a or 100b.
[0039] When speaker 10a speaks, the sound waves from his or her
voice travel to ceiling microphone 20a and are converted to
electrical impulses. Microphone amplifier 30a amplifies these
electrical impulses, and A/D converter 40a converts them to digital
audio data. The digital audio data then travels to the echo
canceller 50a, which taps the output of decoder 70a using
transmission path 130a, to reduce any echo contained in the digital
audio data. Once the echo has been reduced, the digitized audio
data is transferred to the encoder 60a, which encodes the digitized
signal according to a format of the communication link 140. The
communication link 140 then carries the digitized audio data to
room 110b.
[0040] Digital audio data received at room 110a is first decoded by
the decoder 70a according to the transmission protocol of the
communication link 140. This decoded digital audio data is used to
reduce echo, as discussed above, and also converted into electrical
impulses by the D/A converter 80a. The electrical impulses are
amplified by the power amplifier 90a and converted to sound waves
by the loudspeaker 100a.
[0041] Though the above description refers only to room 110a, it is
equally applicable to room 110b. Therefore, the audio portions of
the video teleconferencing systems in rooms 110a and 110b allow
speakers 10a and 10b to simultaneously exchange audio data across
the communication link 140.
[0042] Moreover, microphone amplifier 30a, A/D converter 40a, echo
canceller 50a, encoder 60a, decoder 70a, D/A converter 80a, and
power amplifier 90a may be implemented separately as hardware or
software elements or integrated into a single device such as an
ASIC "System on a Chip". Microphone amplifier 30b, A/D converter
40b, echo canceller 50b, encoder 60b, decoder 70b, D/A converter
80b, and power amplifier 90b may be similarly integrated, or
individually implemented.
[0043] While a video teleconference is described above with respect
to two speakers in two rooms, other configurations are also
possible. For example, three or more rooms may by linked by
communication link 140 to a common teleconference, and more than
one speaker may also be present in each of the rooms. Additionally,
a self-contained, table-top teleconference unit may be used to
allow each speaker to join the teleconference without leaving their
desk, and some speakers may also join the teleconference using
audio-only communications. As those skilled in the art will
recognize, numerous other video teleconferencing configurations are
possible without departing from the scope of the present
disclosure.
[0044] FIG. 2a is an overhead view of room 200 according to an
exemplary embodiment of the present disclosure. Room 200 includes
an exemplary ceiling microphone assembly 210 mounted above an oval
conference table 220. The sensitivity pattern for microphone
assembly 210 includes sensitivity lobe 230 (dashed line), which
define areas of heightened sensitivity. Sensitivity lobe 230 is
aligned with the centre line of the conference table 220, and is
wide enough to cover participants 240 located around the table.
Thus, microphone assembly 210 is more sensitive to sound
originating from participants 240 than from other sources. For
example, microphone assembly 210 is relatively insensitive to sound
coming from the fan 250 and/or reverberant sound 260).
[0045] FIG. 2b is a side view of room 200. As illustrated in FIG.
2b, the sensitivity pattern of the ceiling mounted microphone
assembly 210 is independent of the elevation angle .alpha., and is
highest in an area 270.
[0046] In FIGS. 2a and 2b, the ceiling microphone 210 and table 220
are merely exemplary, and therefore not limiting. Those of skill in
the art will recognize that the microphone assembly 210 can be
mounted at any height and position. The microphone assembly 210 can
also be of any size, form and material that is known.
[0047] Table 220 can also be of any shape, height, and material
that is known, as those of skill in the art will recognize.
Further, though participants 240 are shown positioned around table
220, the participants may also be sitting scattered, i.e. like in a
class room, on rows, i.e. like in a auditorium, or any other
configuration. More than one ceiling microphone assembly may also
be mounted in the same room to cover large areas and room. Multiple
speakers may also be accommodated by microphone assembly 210
without departing from the scope of the invention.
[0048] FIG. 3 is a ceiling microphone assembly 300 according to an
exemplary embodiment of the present invention. Ceiling microphone
assembly 300 includes two plane surfaces 310 and 320
perpendicularly joined to form a structure having an L-shaped cross
section. The two surfaces 310 and 320 preferably form a 90 degree
angle, but other angles are possible without departing from the
scope of the present disclosure. For example, two surfaces 310 and
320 can for angles in the range of 80-100 degrees as those of skill
in the art will recognize. The two surfaces 310 and 320 are made of
a smooth, hard and/or audio reflective surface, such as Plexiglas,
glass, metal and wood.
[0049] The ceiling microphone assembly includes two microphone
elements 330 and 340, such as an omni directional microphone
element, and a subtractor 355. However, as those skilled in the art
will recognize, other types of microphone elements, such as
cardioid and bidirectional are also possible. The output of the
first microphone element 330 is subtracted from the output of the
second microphone element 340 in subtractor 355 and equalized in an
equalizer 370, which has the frequency response of H.sub.eq. The
overall output of the ceiling microphone assembly 300 is the output
of the equalizer 370.
[0050] The first microphone element 330 is arranged substantially
at the intersection between the two surfaces 310 and 320 to capture
both direct sound waves and early reflections from surfaces 310,
320. Preferably the first microphone is arranged in the centre of
the structure formed by the two joined surfaces 310 and 320.
Microphone element 330 is also arranged to exploit the boundary
principle.
[0051] A second microphone element 340 is arranged at a distance
(d) from microphone element 330 along a line bisecting the angle
formed by the two joined surfaces 310, 320. The distance (d) is
preferably chosen so that d 2 is be less than half of a wavelength
of a highest-frequency component to be captured by the ceiling
microphone assembly 300. However, other values for the distance (d)
are possible as will be recognized by those skilled in the art
[0052] In FIG. 3, direct sound waves 380 (solid lines) arrive at
the surfaces 310, 320, and are reflected by one of the surfaces 310
and 320 to form reflected sound waves 390 (dashed lines.) Further
reflection by other surfaces generate reflections 395 (dash-dotted
line). Microphone element 330 captures both the direct sound waves
380 and reflected sound waves 390 from the two surfaces 310 and
320, making use of the boundary principle to increase sensitivity.
Microphone elements 340 receive both direct sound waves 380 and
reflected sound waves 390 that are delayed with respect to the
direct sound waves 380. The amount of delay of the reflected sound
waves 390 depends on the incoming angle (.beta.) and the distance
(d). Any sound waves originating behind and above the ceiling
microphone assembly 300 are blocked by the surfaces 310, 320.
[0053] FIG. 4 is an equivalent diagram of the exemplary ceiling
microphone in FIG. 3. The equivalent diagram mirrors the microphone
elements 330 and 340 about each surface 310 and 320, and the
surfaces 310 and 320 are removed. In other words, the equivalent
diagram of FIG. 4 includes five microphone elements implementing a
ceiling microphone functionally equivalent to the ceiling
microphone in FIG. 3.
[0054] More specifically, in FIG. 4 second microphone 340 is first
mirrored around the first surface 310 and then the second
microphone 340 and its mirrored equivalent 340b is mirrored around
the second surface 320 to generate two additional mirrored
microphone elements 340c and 340d. The total output of the second
microphone element 340 equals the sum of the four equivalent
microphones in FIG. 4 (340, 340b, 340c and 340d). The result is
four versions of the same audio signal with four time delays. As
those of skill in the art will recognize, the time delays
corresponding to the four microphone elements 340, 340b, 340c and
340d may be different from one another, or two or more of the time
delays may be the same.
[0055] Mirroring the first microphone element 330 around each of
the two surfaces 310, 320 also results in four equivalent
microphones. However, as the first microphone 330 is located at the
vertex of the angle formed by surfaces 310 and 320, all of the
equivalent microphones coincide to the same location as the first
microphone element 330. As there is no relative delay among the
four equivalent microphones corresponding to the first microphone
element 330, signal level is four times that of the first
microphone signal 330. This is the exact same phenomena as the
pressure quadrupling caused by two surfaces.
[0056] FIG. 5 is an equivalent circuit corresponding to the
equivalent diagram of FIG. 4. In FIG. 5, the first microphone
element 330 outputs a first signal corresponding to the incoming
sound wave acoustically amplified by a factor of four due to the
two surfaces 310 and 320. The second microphone element 340 outputs
a sum of four delayed versions of the same signal. The ceiling
microphone of FIGS. 3, 4 and 5 therefore implements the same
directive pattern for all signals impinging upon the ceiling
microphone from an angle .alpha. in between the two surfaces 310
and 320.
[0057] Thus, the directivity pattern of the microphone assembly 300
is independent of the elevation angle alpha. Instead, the
directivity pattern approximates g(.beta.)=cos.sup.2 (.beta.),
where .beta. is the angle between the incoming sound and a line
defined by the intersection of the two surfaces 310, 320. The
resulting directivity pattern is a quarter of a second order
toroidal pattern of FIG. 6. To obtain this pattern, an equalizing
filter H.sub.eq(.omega.) must be proportional to 1/.omega..sup.2
for obtaining a flat frequency response.
[0058] Sound waves captured by microphone elements 330 and 340 are
converted to electronic signals thereby and the signal from the
first microphone element 330 is subtracted from the signal from the
second microphone element 340. In FIG. 3, the signal from the first
microphone element 330 is inverted with a signal inverter 350 and
subsequently added to the signal from the second microphone element
340 in an adder 360. Such a configuration can, for example, be
implemented as a purely analogue system. Alternatively, subtractor
355 can be only of an adder (not shown).
[0059] In a digital system, an adder circuit may also be used as
subtractor 355. For example, subtractor 355 can be any unit able to
subtract two signals, such as those from the first and second
microphone elements 330, 340. The output of subtractor 355 is then
equalized by equalizer 370, which has a frequency response
(H.sub.eq) of:
H eq ( .omega. ) = 1 .omega. 2 , ##EQU00001##
where w is the frequency in radians per second.
[0060] The gain of inverter 350, equalizer 370 and adding node 360
may be implemented as digital structures, in which case A/D
converters (not shown) convert the analog electrical impulses from
microphone elements 330, 340 into digital audio data. Equalizer 370
can be implemented as infinite impulse response (IIR) filters or
finite impulse response (FIR) filters.
[0061] Subtractor 355, inverter 350 and equalizer 370 may also be
implemented separately or integrated in a single device. For
example, the Subtractor 355 and equalizer 370 may be implemented on
a PC computer 400, such as the one in FIG. 10. The computer 400
includes a processor 405 for performing computations, a read-only
memory (ROM) 430 for storing programming instructions, and a main
memory 425 that may include RAM memory, FLASH memory, EEPROM memory
or any other known rewritable memory. The main memory 425 stores
temporary data, instructions, etc. The computer 400 also includes a
display controller 420 for controlling a display device 460, a disk
controller 435 for controlling a hard disk 445 and/or a CD-ROM
drive 440, and an I/O interface 410 for controlling a pointing
device 450 and a keyboard 455. A bus 415 interconnects all of the
above-described components.
[0062] Hard disk drive 445 and CD-ROM drive 440 may be integrated
into the computer 400, or may be removable. Likewise, at least a
portion of the main memory 425 may also be removable. Though not
shown in FIG. 10, the I/O interface 410 may also interface to a
network, phone system, WiFi network, cellular network, WAN, LAN,
etc.
[0063] Subtractor 355, equalizer 370 and inverter 350 may also be
implemented on computer 400 as a utility application, background
demon, or component of an operating system, or any combination
thereof executing in conjunction with the processor 405 and an
operating system, such as Microsoft VISTA, UNIX, SOLARIS, LINUX,
Apple MAC-OS and other systems known to those skilled in the
art.
[0064] Subtractor 355, inverter 350 and equalizer 370 may be
implemented in hardware, together or separately, on devices such as
FPGA's, ASIC's, microcontrollers, PLD's, or other computer readable
media such as an optical disc.
[0065] In addition, microphone noise from microphone elements 330,
340 may be mitigated using bandpass filters to filter the signals
from each of the microphone elements 330, 340. For example, such
band pass filters can have a high-pass roll off frequency of 80
hertz since attenuation of frequencies below 80 hertz minimally
impacts sound quality, but reduces noise levels attributable to the
microphones 330 and 340, A/D converter, quantization and/or
numerical rounding.
[0066] Alternatively, the band pass filters may have different
highpass roll-offs. For example, the second band pass filter may
have a higher high-pass roll off frequency than the first band pass
filter so that the signals generated by adding node 360 (or
subtractor node) include only signals from the first microphone 330
element for low frequencies. This degrades the directivity pattern
at low frequencies, but also reduces system noise.
[0067] A degradation of the directivity pattern at high frequencies
is also acceptable in order to reduce system noise. Increasing the
distance d between the second microphone element 340 and the
surfaces 310, 320 causes the microphone assembly 300 to have higher
sensitivity to low frequencies. This may cause some spatial
aliasing at high frequencies, but also reduces system noise.
[0068] Alternatively, the system may only use the first microphone
330 for higher frequencies (as described for low frequencies in the
previous paragraph), resulting in an omni directional response at
high frequencies. As air dissipation dampens high-frequency
reverberations, omni directional high frequency responses still
yield acceptable overall sound quality.
[0069] FIG. 7 is another exemplary embodiment of the ceiling
microphone assembly according to the present disclosure. In FIG. 7
three omni directional microphone elements 330, 340 and 335 are
used to reduce the impact of system noise. Microphone element 330
is placed substantially at the intersection between the two
surfaces, and microphone element 340 and 335 are aligned to
microphone 330 with respect to a line bisecting the angle formed by
the two joined surfaces 310, 320 to capture both direct sound waves
and sound waves reflected by the surfaces 310, 320. Preferably, the
first microphone 330 is arranged in the centre of the structure
formed by the two joined surfaces 310, 320. Microphone 340 is a
distance (d) from both surfaces 310 and 320, and microphone 335 is
twice the distance (d) from both surfaces 310 and 320. In the above
description, the distance (d) is preferably chosen such that d 2 is
less than half of a wavelength of a highest-frequency component to
be captured by the ceiling microphone assembly 300. However, as
those of skill in the art will recognize other values of (d) may be
used without departing from the scope of the present invention.
[0070] The sound waves captured by microphone elements 330, 340 and
335 are converted into electronic signals thereby and combined in
low and high frequency subtractors 770 and 775. Specifically, the
signal from the first microphone element 330 is subtracted from the
signal of the third microphone element 350 in low frequency
subtractor 770. The output of the low frequency subtractor 770 is
then equalized by low frequency equalizer 780 and filtered by low
pass filter 785.
[0071] For high frequencies, the signal from the first microphone
element 330 is subtracted from the signal of the second microphone
element 340 in the high frequency subtractor 775. The output of the
high frequency subtractor 775 is then equalized by high frequency
equalizer 790 and high pass-filtered by high pass filter 795.
[0072] The outputs of low pass filter 785 and high pass filter 795
are summed at the summing node 799 to obtain the output of the
ceiling microphone assembly.
[0073] As discussed above, the subtractors 770 and 775 can be
implemented with a signal inverter and adder, only with an adder
(as shown), or any other unit able to subtract signals that is
known. Further, subtractors 770 and 775 may be implemented as
analog or digital circuits. Also in FIG. 7
HH.sub.eg=HL.sub.eg=H.sub.eq.
The high pass filter 795 removes any low frequency components
remaining in the output of equalizer 790, and low pass filter 785
removes any remaining high frequency components before being summed
at summing node 799 to generate the overall ceiling microphone
output. Thus, the ceiling microphone assembly of FIG. 7 uses
microphones 330 and 340, which are closely spaced together, to
capture high-frequency sound waves, and microphones 340 and 335,
which are spaced further apart, to capture low-frequency sound
waves. This two-way system implements a high frequency quarter
toroid sensitivity pattern and a low frequency quarter toroid
sensitivity pattern to remove system noise without distorting
microphone sensitivity. As those skilled in the art will recognize,
the two-way system of FIG. 7 may be extended to a three-way system,
four-way system, or even an n-way system, where n is any positive
integer. Further, any of the above-described system noise reduction
techniques may be combined to further optimize performance of the
ceiling microphone assembly.
[0074] FIG. 8 is a further exemplary embodiment of the ceiling
microphone assembly according to the present disclosure. The
ceiling microphone in FIG. 8 is implemented using one bidirectional
microphone and two waveguides (e.g. tubes). In FIG. 8, a
bidirectional microphone 830 is positioned approximately at a
distance d/2 from each of the two surfaces 310 and 320. The
bidirectional microphone 830 has a front and a rear acoustical
input port for allowing sound to enter the microphone from opposite
sides. A first waveguide 850 (or tube) have a first waveguide
output port that is connected to (associated with) the rear
acoustic input port of the bidirectional microphone 830. The first
waveguide's 850 input port is arranged adjacent to the intersection
of the first and second surfaces 310, 320. A second waveguide 840
has an output port associated with the front acoustic input port of
the bidirectional microphone 830. The second waveguide input port
is arranged a predetermined distance (d) from the first and second
surfaces 310, 320.
[0075] Each waveguide 840, 850 may be any linear structure that
guides electromagnetic waves. The first and second waveguides 840,
850 are of equal dimensions (length, width, height) and probe audio
pressure. The first waveguide 850 probes the audio pressure from
the corner between the surfaces 310, 320 and the first waveguide
840 probes the audio pressure at a point displaced by d both
horizontally and vertically from the corner. The waveguides 840,
850 transfer the air pressure to the opposite sides of a
bidirectional microphone's 830 membrane. Since the two pressures
enter at different sides of the membrane, a subtraction function
between the two pressures is implemented. In FIG. 8, the equalizing
filter H.sub.eq(.omega.) 860 includes a 1/.omega..sup.2 factor, and
also takes into account any frequency dependency caused by the
tubes 840, 850. Such dependencies depend on both the length and
width of the tube 840 or 850, as well as the bidirectional
microphone 830 itself. The tubes 840 and 850 are preferably equal
on both sides of the bidirectional microphone 830 for proper
performance.
[0076] In addition, the 1/4-wavelength resonances of the tubes 840,
850 set a upper frequency limit, and define the size of and
distance to the reflecting surfaces 310, 320. The concept of
Fresnel zones can be used to estimate when a surface is big enough
to be considered a reflector at a certain distance l, as those of
skill in the art will recognize. Assuming plane waves, the
relationship is then given by
fa.sup.2=pl,
where f is the frequency, a is the smallest dimension of the
surface and p is a proportionality constant.
[0077] Though the above discussion has been made with reference to
traditional microphone elements, other microphone elements may also
be used without departing from the spirit of this disclosure. For
example, optical microphones and/or MEMs microphones may be used.
Optical microphone may reduce the discussed noise problems
dramatically. MEMs microphones have the advantage of allowing
better component matching if all components, including the
microphone, are fabricated on the same silicon wafer or the same
silicon die. Of course, the equalizer filter response may have to
be modified accordingly. Using the technique with one bi
directional microphone with tubes, match between microphones is no
longer an issue, since this is implemented by using one single
microphone/microphone membrane. Match of the tubes are important,
but easily realized.
[0078] As first recognized by the present inventor, the ceiling
microphone assembly described herein is mounted separately from a
conference table. Therefore, it has relatively low sensitivity with
respect to audio originating from the table (paper shuffling,
keyboard noise from laptops, etc). The ceiling microphone assembly
also has a "line of sight" for direct sound from the parties to the
teleconference, regardless of any PCs or similar obstructions that
may be situated on the conference table. As the ceiling microphone
assembly is mounted from the ceiling, no cables are present on the
table. Further, the ceiling microphone is fixed and therefore is
not vulnerable to incorrect use or displacement.
[0079] As first recognized by the present inventors, the ceiling
microphone assembly according to the present disclosure introduces
directivity to double surface boundary ceiling microphones,
reducing reverberation and extending reach. The ceiling microphone
described herein has a directivity pattern better suited for normal
oval-shaped conference tables, and can be implemented using one
microphone, eliminating the need for calibration of microphone
elements.
[0080] Obviously, numerous modifications and variations of the
present invention are possible in light of the above teachings. It
is therefore to be understood that within the scope of the appended
claims, the invention may be practiced otherwise than as
specifically described herein.
* * * * *